Y.-M. Ahn, R. F. Pratt / Bioorg. Med. Chem. 12 (2004) 1537–1542
1541
Arg 346, which are available in the b-lactamase to
orient the heterocyclic ring of a bicyclic b-lactam
away fromthe acyl group and thus allow deacylation,
may be a significant factor in the inhibition of this
enzyme by b-lactams.
tion (2 atm) over 10% Pd/C in ethanol. The product, 6,
was recrystallized from4/1 benzene/ethyl acetate. Melt-
ꢃ
1
ing point 196–198 C. H NMR (2H6-DMSO) d 7.5–7.7
(m, 6H, ArH), 7.8–7.9 (m, 3H, ArH), 7.95 (d, J=8 Hz,
2H, ArH), 8.26 (d, J=8 Hz, 2H, ArH).
The principal conclusion fromthe experiments described
above is that the phenylphosphate leaving group is
much more effective than the m-hydroxybenzoate and
thioglycolate groups in promoting acylation of the P99
b-lactamase. This probably arises from the optimal
placing of negatively charged oxygens in the active site
to interact with Tyr 150 and Lys 315. Tyr 150 may be
able to assist departure of the leaving group by hydrogen-
bonding/proton transfer to the non-bridgehead phos-
phoryl oxygen as well as it can by proton donation to
the bridgehead (departing) oxygen. Design of inhibitors,
both covalent and non-covalent, for class C b-lactamases,
should take into account the optimal positioning of
anions in the active site.
The benzoyl analogue, 4, was prepared in the same way
and recrystallized from3/2 ethyl acetate/hexane. Melting
point 188–190 ꢃC. 1H NMR (2H6-DMSO) d 7.6 (m, 4H,
ArH), 7.76 (t, J=7 Hz, 1H, ArH), 7.80 (s, 1H, ArH),
7.88 (d, J=7 Hz, 1H, ArH), 8.14 (d, J=7 Hz, 2H,
ArH).
3.3. Analytical and kinetics methods
All kinetics experiments were performed at 25 ꢃC in 20
mM MOPS buffer, pH 7.5. The concentrations of stock
solutions of the enzyme were obtained spectrophoto-
metrically. Reactions of 4–7 were monitored spectro-
photometrically at 318 nm (Áe=ꢀ8130 Mꢀ1 cmꢀ1), 318
nm( Áe=ꢀ225 Mꢀ1 cmꢀ1), 312 nm( Áe=ꢀ3450 Mꢀ1
cmꢀ1), and 346 nm( Áe=ꢀ166 Mꢀ1 cmꢀ1), respectively.
The accessible concentrations of 4–7 were limited by
solubility such that the highest concentrations of each in
the kinetics runs were 0.7 mM, 5.0 mM, 0.8 mM, and
5.0 mM, respectively. Steady state kinetics parameters
were obtained frominitial rates by non-linear least
squares fitting of the data to the Michaelis–Menten
equation. Enzyme-catalyzed methanolysis experiments
were performed as described previously;3 methanol
concentrations were varied fromzero to 2.5 M. Sub-
strate aminolysis kinetics were also conducted as
described in a previous paper;7 d-alanine and d-phenyl-
alanine were employed with concentrations ranging up
to 200 mM and 48 mM, respectively.
3. Experimental
The Enterobacter cloacae P99 b-lactamase was pur-
chased fromthe Centre for Applied Microbiology and
Research, Porton Down, Wilts., UK, and used as
received.
3.1. 4-Phenylbenzoylthioglycolic acid (7)
4-Phenylbenzoic acid (Aldrich; 3.63 g, 18.3 mmol) was
dissolved in dry tetrahydrofuran (80 mL) with stirring
under a nitrogen atmosphere and cooled to 0 ꢃC. Solid
carbonyl diimidazole (Aldrich, 3.0 g., 18.5 mmol) was
added and the mixture stirred for 1 h. Mercaptoacetic
acid (Aldrich 1.28 mL, 18.3 mmol), dissolved in tetra-
hydrofuran (15 mL), was added and the reaction mixture
stirred at 0 ꢃC for four days. After rotary evaporation of
the solvent, the residue was taken up into ethyl acetate
(100 mL) and washed with 10% citric acid (2ꢁ100 mL)
and water (2ꢁ100 mL). After drying over MgSO4, the
solvent was removed by rotary evaporation and the
product recrystallized from4/1 benzene/ethyl acetate in
40% recovered yield. Melting point 170–172 ꢃC. 1H
NMR (2H6-DMSO) d 3.93 (s, 2H, CH2), 7.4–7.6 (m,
3H, ArH), 7.7–7.8 (m, 2H, ArH), 7.89 (d, J=8.5 Hz,
2H, ArH), 8.02 (d, J=8.5 Hz, 2H, ArH).
The structures of Figure 1 were derived fromcomputa-
tional models of enzyme–substrate complexes that were
set up essentially as previously described5,7 and run on
an SGI Octane 2 computer with INSIGHT II (MSI, San
Diego, CA). The starting point was the crystal structure
of the P99 b-lactamase with a phosphonate inhibitor
covalently attached to the active site serine residue
(PDB file 1 bls18). This was transformed into the acyl-
ation tetrahedral intermediate 11 by means of the
builder module of INSIGHT II. In this model, Lys 67
and Lys 315 were cationic, Tyr 150 was neutral, and the
tetrahedral intermediate 11 was dianionic. A variety of
initial conformations of the ligand were explored by
manual positioning followed by molecular dynamics.
Two general conformations seemed to dominate (see
text and Fig. 1). Those were not interconvertable in 200
psec dynamics runs (where the entire protein together
with solvating water molecules were unrestricted).
Typical snapshots of each of these conformations were
selected for energy minimization. The minimized struc-
tures were then used to calculate interaction energies,
Eint,19 between ligand and protein. The residues inclu-
ded in these calculations were Ser 64, Lys 67, Tyr 150,
Asn 152, Lys 315, Thr 316, Gly 317, and Thr 318, the
residues shown to best discriminate between tetrahedral
ligands,20 and also Arg 349 which was found to interact
with the carboxylate of the leaving group (see text).
The benzoyl analogue, 5, was prepared in the same way
in 54% yield after recrystallization from4/1 benzene/
ꢃ
1
ethyl acetate. Melting point 94–96 C. H NMR (2H6-
DMSO) d 3.89 (s, 2H, CH2), 7.59 (t, J=7.5 Hz, 2H,
ArH), 7.72 (t, J=7.5 Hz, 1H, ArH), 7.95 (d, J=7.5 Hz,
2H, ArH).
3.2. 3-[40-phenylbenzoyloxy]benzoic acid (6)
4-Phenylbenzoic acid was condensed with benzyl 3-
hydroxybenzoate in the presence of carbonyl diimida-
zole as described above. After recrystallization from
cyclohexane the benzyl ester of 6 was obtained in 57%
yield. Deprotection was then performed by hydrogena-